How do i monitor zabbix itself?

[tiv]

i just set up zabbix to monitor my three servers and it's working great. one question i have though is that if the server that zabbix is running on dies, zabbix wont notify me since it wont be running. what can i do about this?

[xs-]

What was first, the chicken or the egg?
IMHO: monitoring is a secondary process, never a critical business app (like backup), if it fails by itself, it shouldn't matter for your critical processes.

I'd search freshmeat for some simple process checking + action triggering tool which can send you an email or sms.

[just2blue4u]

you could check if the server that's running zabbix is listening on port 10051 (standard zabbix server port).

or you EnableRemoteCommands=1 in agent config and let the server create a file on that agent every minute, then let the client check if the file is older than 1 minute.

Best solution would be zabbix distributed monitoring (with 2 zabbix servers). But i don't think you really wanna do that.

[Tenzer]

I have monit running on my Zabbix server, which automatically starts the zabbix_server process again in case it dies. It keeps track of how often it restarts the process, and if it goes over a certain threshold it will stop trying. It informs about all this by sending e-mails, so the only thing it won't notice, is if the Zabbix servers's hardware goes down, but I have another Zabbix server to monitor that

[ldil]

Hi tiv

I did this, running gnokki with a gsm phone and looking at port 10050 and 10051 on the local machine. If it fails send a mail and sms with sendsms/gnokki.

You can also do this with Webmin monitoring.

Greetz Leo

[tiv]

so, it seems like setting up another zabbix for distributed monitoring might be overkill when we juts have 3 servers running total.

ldil, what is this webmin monitoring?

-v

[tore]

We had some problems with the zabbix server process dieing, so I made a small bash script (running on the zabbix server) that restarts zabbix if it has died. It also writes a log so you can keep track of any restarts. Here is the code:

#!/bin/bash
log=/var/log/zabbix/zabbix-autorestart.log
procno=`ps -eF | grep -c "[z]abbix_server"`
if (( $procno == 0 )); then
# DEBUG echo "Zabbix not OK, RESTART!"
	/etc/init.d/zabbix-server stop
	/etc/init.d/zabbix-server start
	echo `date` >> $log
# DEBUG else 
# DEBUG echo "Zabbix OK!"
fi
exit 0

I run this script every minute via crontab

--
Tore